Citation

Derouen KR, Bercovitch J. J. Peace Res. 2008; 45(1): 55-74.

Copyright

(Copyright © 2008, SAGE Publishing)

DOI

10.1177/0022343307084923

PMID

unavailable

Abstract

The enduring rivalry (ER) framework was developed for studying the long-term dynamics of serious conflicts between pairs of states. Here, the logic and structure of that framework is applied to civil wars. Many civil wars are very long and recur often. A new way of thinking of about these long and seemingly interminable internal conflicts emphasizes a dyadic perspective and enduring internal rivalry (EIR). Within this framework, the article demonstrates empirically that EIRs are different from other wars. This study offers a definition and an initial dataset of EIRs. Working from the Uppsala Conflict Termination Dataset, we find that about 76% of all civil war years from 1946 to 2004 took place in the context of EIRs. Several statistical models are tested to demonstrate the empirical validity of the EIR construct, while controlling for state capacity, democracy, type of termination, military coups, war intensity, and duration of war. The logit results provide evidence that civil wars involving EIRs are more likely to recur, and the hazard analysis results reveal that EIRs are followed by shorter peace spells. The early phases of EIRs are followed by relatively shorter peace spells, thus indicating a `locking in' period that scholars have identified in international rivalries. Military victories lead to longer peace, but few EIRs are terminated with military outcomes. The hazard models employed here employ repeat-event techniques, since many civil wars exhibit patterns of recurrence. Implications of these results for conflict management are offered.